KCHS Math Presentation

Nehemias Ulloa

About Me

  • Bakersfield, CA

About Me

  • Bakersfield, CA
  • CSU Bakersfield - BS Math with Statistics Emphasis

Ariel View of CSUB Campus

CSUB Roadrunner

About Me

  • Bakersfield, CA
  • CSU Bakersfield - BS Math with Statistics Emphasis
  • Iowa State University - Ph.D. Statistics

Iowa State Campus

What Do I Do

  • CVB (Center for Veterinary Biologics)
    • USDA - MRP - APHIS - VS
    • License Biologics
      • Regulate Vaccines
      • Regulate Assays
      • Regulate Diagnostics

Regulating Vaccines

  • Company wants to license vaccine
    • e.g. Bordetella bronchiseptica Vaccine in Dogs
    • Submit a Licensing Plan
    • Submit Multiple Studies
      • Safety Study
      • Shed/Spread Study
      • Assay Validation
      • Efficacy Study

Example - Lung Lesions in Dogs

  • Study Design

    • 2 Treatments Groups
    • Randomization
    • Blinding
    • Vaccinated/Challenged on Same Day
    • Similar Housing/Living Conditions
    • Same Breed
    • Both Male and Female Dogs

Boxplot

mlesion %>%
  ggplot(aes(y=les, x=tx, color=tx)) +
    geom_boxplot() + 
    geom_point() +
    labs(y="Lung Lesion Score", color="Treatment", x="Treatment") + 
    scale_color_manual(values=c("black", "red")) +
    theme_bw()

Boxplot

Dot Plot

mlesion %>%
  ggplot(aes(x=les, y=animalID, color=tx)) +
    geom_point() + 
    labs(x="Lung Lesion Score", y="AnimalID", color="Treatment") + 
    scale_color_manual(values=c("black", "red")) +
    theme_bw()

Dot Plot

Analysis

Mean

mlesion %>%
  group_by(tx) %>%
  summarise(`Mean Lung Lesion Score` = mean(les))
# A tibble: 2 × 2
  tx    `Mean Lung Lesion Score`
  <fct>                    <dbl>
1 con                       8.08
2 vac                       3.54

Standard Deviation

mlesion %>%
  group_by(tx) %>%
  summarise(`Lung Lesion Score sd` = sd(les))
# A tibble: 2 × 2
  tx    `Lung Lesion Score sd`
  <fct>                  <dbl>
1 con                     8.79
2 vac                     4.46

Median

mlesion %>%
  group_by(tx) %>%
  summarise(`Median Lung Lesion` = quantile(les, probs=0.5))
# A tibble: 2 × 2
  tx    `Median Lung Lesion`
  <fct>                <dbl>
1 con                   5   
2 vac                   3.45

Analysis

mlesion %>%
  group_by(tx) %>%
  summarise(`Mean Lung Lesion Score` = mean(les),
            `Lung Lesion Score sd` = sd(les),
            `Median Lung Lesion` = quantile(les, probs=0.5)) %>%
  mutate(across(where(is.numeric), round, 2))
# A tibble: 2 × 4
  tx    `Mean Lung Lesion Score` `Lung Lesion Score sd` `Median Lung Lesion`
  <fct>                    <dbl>                  <dbl>                <dbl>
1 con                       8.08                   8.79                 5   
2 vac                       3.54                   4.46                 3.45

Analysis

Analysis

  • Mitigated Fraction
    • MF is the relative increase in the probability that a vaccinate’s disease will be less severe than a nonvaccinate’s disease.

    • MF evaluates the intervention’s effect by the probability a subject will benefit from the intervention

    • Here’s the code

      MF::MFr(les ~ tx, data=mlesion)
      [1] 0.2011834

Analysis

What if we knew more information about the housing?

Analysis - Matched Pairs

MF::MFmp(les ~ tx + cluster(cage), data=mlesion)
95% t intervals on 24 df
 
    point     lower     upper 
0.3846154 0.1316679 0.6375628 

Finally

Questions?